Prototyping and Transforming Facial Textures for Perception Research
IEEE Computer Graphics and Applications
Non-negative Matrix Factorization with Sparseness Constraints
The Journal of Machine Learning Research
Hi-index | 0.00 |
As far as the majority of known aging methods are concerned, PCA (Principal Component Analysis) was used as the first step to extract facial features and build model space. In this paper, NMF (Non-negative Factorization) with sparseness constraints is used as an alternative to PCA in the feature extraction step when aging an unseen human face image to the required age. A variety of experiments demonstrate that by adding sparseness constraints to NMF we can get simulated aging faces which share more similarities with real images than those by the method of PCA, especially when we keep the coefficients sparse while leaving the basis vectors unconstrained.